bumbling workerbee

How can we predict performance when load testing a server or set of servers?

q chat If you run more than one q chat in tmux, you can have it manipulate your other panes and windows, effectively giving you a communication bus.

I find myself giving a version of these definitions semi-regularly now to people outside of “the space,” and don’t really check them against other definitions from inside “the space.”

If you Google “adding replicas can’t improve search latency elasticsearch”, you’ll find Distinguished Software Engineer Adrien Grand saying something that should be true: Replicas do not help with latency, they can only help with throughput by replicating…

There’s a now famous rant on async I read some years ago before I worked with async much.

There’s many tools engineers underuse, but I want to pick on three I don’t see discussed often: dates, affect, and agency.

Whenever possible, it’s nice to know the theoretical limits of what you’re doing.

HNSW, which stands for “Hierarchical Navigable Small World” (graphs), is one of the first data structures for which I read not just its paper but (a lot of) the trail of papers backward in time until I arrived at one with a name I already knew. the goal a…

On every team I’ve worked on, including at Amazon, engineers bring up “tech debt” as a concern.

In day to day work I don’t get to make much use of dynamic programming.